An image-processing program and an image processor well known in the art employ a method for converting multi-value image data to fewer-multi-value image data having fewer tone gradations than the multi-value image data, for example, binary image data. One such image-processing method is disclosed in Japanese unexamined patent application publication No. 2002-125122. According to this image-processing method, the image-processing program and image processor employ dither data in order to convert image data to be outputted to the printer to image data having binary image data than the multi-value image data being processed.
In the dither data used in the above image-processing method, the threshold values compared to the densities of pixels belonging to each block are arranged so as to increase the uniformity of dot density. Hence, when converting multi-value image data having a relatively low density, patterns formed by dots in the image data change irregularly when the targeted multi-value image data is converted to low-level data. Consequently, the patterns in images formed by the printer also change irregularly. In other words, converting multi-value image data having a relatively low density produces noticeable irregular patterns in images formed on a printer, particularly in monochromatic images.